SCIENTIFIC ARTICLE • SEQUENCING & BIOINFORMATICS

Reference-Based vs De Novo Analysis of Sequencing Data

A scientific overview of two fundamental strategies for analyzing sequencing data, exploring how reference-based mapping and de novo approaches differ, when each method is appropriate, and how they can complement one another in modern genomic research.

→ BIOINFORMATICS · GENOME ANALYSIS

Start Reading ↓

01 / INTRODUCTION

Reference-Based vs De Novo Analysis of Sequencing Data

Understanding how sequencing reads are transformed into meaningful genomic information starts with one fundamental question: should the data be analyzed against an existing reference, or assembled without one?

Modern sequencing technologies can generate large numbers of DNA reads, but these sequences do not automatically provide a complete representation of a genome. Computational methods are therefore required to organize, compare, or reconstruct the information contained within sequencing data.

Two major strategies are commonly used in genomic analysis. In a reference-based approach, sequencing reads are aligned or analyzed using a previously available reference genome as a guide. In a de novo approach, the sequence is reconstructed directly from sequencing reads without using a previously resolved genome as the assembly template.

These approaches answer related but different biological questions. Reference-based analysis can provide an efficient framework for studying samples from organisms with suitable reference genomes, whereas de novo assembly can help recover sequence information that may be absent, divergent, or poorly represented in an existing reference.

TWO ANALYTICAL STRATEGIES

Where does the reference genome fit?

REFERENCE-BASED

Guided by an existing genome

Sequencing reads are mapped or analyzed using a previously available reference sequence as a guide.

Sequencing reads → Reference genome → Analysis

DE NOVO

Reconstructed from sequencing data

Sequencing reads are assembled without using a previously resolved genome as the assembly template.

Sequencing reads → Assembly → Genome sequence

WHY IT MATTERS

The choice of strategy can influence what the analysis reveals

A reference genome provides a coordinate system that can make many analyses efficient and straightforward. However, reliance on a reference can also limit the ability to characterize sequences that differ substantially from it or are not represented in it. De novo assembly takes a different approach by attempting to reconstruct the underlying sequence directly from the reads.

SCIENTIFIC SOURCE Next-Generation Sequence Assembly: Four Stages of Data Processing and Computational Challenges BioMed Research International

Read scientific article

02 / REFERENCE-BASED ANALYSIS

What Is Reference-Based Analysis?

A reference-based approach uses a previously assembled genome sequence as a coordinate framework for interpreting newly generated sequencing reads.

DEFINITION

Using a known genome as an analytical framework

In reference-based sequencing analysis, reads generated from a biological sample are compared with a previously available reference genome. Computational alignment algorithms determine where individual reads best match the reference sequence.

Once the reads have been aligned, the resulting alignment data can be used for downstream analyses such as variant identification, genotype comparison, coverage assessment, and other forms of genomic interpretation.

ANALYTICAL WORKFLOW

From sequencing reads to genomic interpretation

REFERENCE-BASED

Sequencing

DNA is sequenced to generate a collection of raw sequencing reads.

→

Alignment

Reads are aligned against a selected reference genome.

→

Alignment Data

The aligned reads can be stored and inspected using formats such as SAM or BAM.

→

Interpretation

Researchers investigate variants, coverage and other biological signals within the aligned data.

THE REFERENCE GENOME

Think of the reference as a genomic coordinate system

A reference genome provides an ordered sequence against which reads can be positioned. This makes it possible to describe genomic observations relative to known locations rather than reconstructing the entire genome for every sample.

REFERENCE

ATGCTACG G CTTAGGCA

READ

Reads are positioned relative to the reference sequence

ADVANTAGES +

01 Efficient for organisms with a suitable reference genome
02 Provides a consistent genomic coordinate system
03 Well suited to many resequencing and variant-analysis workflows
04 Generally less demanding than reconstructing an entire genome de novo

LIMITATIONS −

01 Requires an appropriate reference sequence
02 Highly divergent or novel regions may be difficult to interpret
03 Reference-related bias can influence some analyses
04 Structural variation can be challenging to capture completely with a linear reference

SCIENTIFIC SOURCE Whole genome sequencing Review · Nature Reviews Genetics

Read PubMed article

03 / DE NOVO ANALYSIS

What Is De Novo Analysis?

De novo analysis reconstructs genomic sequence directly from sequencing reads, without using a previously assembled genome as the primary assembly template.

DEFINITION

Reconstructing sequence from the reads themselves

In a de novo genome assembly, sequencing reads are computationally combined according to sequence information contained within the reads themselves. The objective is to reconstruct longer stretches of DNA without requiring a previously resolved reference genome.

The resulting assembly is commonly represented first as contigs, continuous assembled sequences. Additional information, such as paired-end, mate-pair, long-read, or other linking evidence, can then be used to order and orient contigs into larger structures called scaffolds.

ASSEMBLY CONCEPT

From fragmented reads to assembled sequence

DE NOVO

Sequencing reads

The sequencing instrument produces many individual DNA reads originating from different regions of the genome.

ATGCTA GCTAAC CTAACC ACCGTA GTACGG TACGGA

Sequence relationships

Assembly algorithms identify sequence relationships between reads or smaller sequence units such as k-mers, depending on the assembly strategy.

ATGCTA

GCTAAC

TAACCG

Contig construction

Consistent sequence information is combined into longer continuous sequences called contigs.

CONTIG

ATGCTAACCGTACGGA

Scaffolding

Additional long-range information can help order and orient contigs into scaffolds, potentially spanning unresolved regions.

CONTIG

NNN

CONTIG

ASSEMBLY ALGORITHMS

Different algorithms solve the assembly problem in different ways

The characteristics of the sequencing reads influence the assembly strategy. Short-read assemblers commonly use de Bruijn graph approaches, whereas many long-read assemblers use overlap-based strategies such as Overlap–Layout–Consensus.

SHORT-READ ASSEMBLY

De Bruijn Graphs

Reads are broken into fixed-length sequences called k-mers. These sequence units are represented as a graph that can be traversed to reconstruct longer sequences.

Learn about de Bruijn graphs

LONG-READ ASSEMBLY

Overlap–Layout–Consensus

Reads are compared to identify overlaps, arranged into layouts, and then used to derive consensus sequences. This approach is widely associated with long-read assembly.

Read about OLC assembly

WHY USE DE NOVO?

De novo assembly can reveal sequence beyond the limits of an existing reference

A reference-based analysis is inherently anchored to the sequence represented by the reference. De novo assembly provides an alternative for organisms without an adequate reference and can help characterize genomic regions that are difficult to represent through mapping alone.

This makes de novo assembly particularly valuable for non-model organisms, genome discovery projects, and studies where structural or highly divergent sequence is of biological interest.

IMPORTANT

De novo does not mean automatically complete

Assembly quality depends on factors including sequencing read length, accuracy, coverage, genome complexity, repetitive regions, heterozygosity, and the assembly algorithm. A de novo workflow therefore produces an assembly that must be evaluated for quality and completeness.

04 / COMPARISON

Reference-Based vs De Novo

The two approaches answer different computational and biological questions. Understanding their differences is essential when designing a sequencing analysis workflow.

01 REFERENCE-BASED

Where do my reads fit within a known genome?

Sequencing reads are aligned against an existing reference sequence to determine their genomic location and support downstream analysis.

02 DE NOVO

What sequence can be reconstructed from my reads?

Sequencing reads are assembled into longer sequences without using an existing genome as the primary assembly template.

SIDE-BY-SIDE ANALYSIS

Key differences

SELECT A CRITERION

REFERENCE-BASED

Mapping against known sequence

REFERENCE REQUIREMENT

Requires a suitable reference sequence against which reads can be aligned.

DE NOVO

Reconstructing sequence

REFERENCE REQUIREMENT

Does not require an existing reference genome for the initial assembly.

DECISION MATRIX

Which approach fits the question?

There is no universally superior strategy. The appropriate approach depends on the biological system, available reference, sequencing technology, and analytical objective.

Criterion	Reference-Based	De Novo
Existing reference	Preferred Requires a suitable reference for mapping.	Not required Assembly can start from sequencing reads.
Primary question	Locate and interpret reads relative to known sequence.	Reconstruct genomic sequence from the reads.
Novel sequence	Limited Sequence absent from the reference may be difficult to recover through mapping.	Stronger option Can reconstruct sequence without requiring a predefined reference coordinate system.
Computational workflow	Generally focused on read alignment and downstream interpretation.	Requires genome assembly and assembly-quality evaluation.
Variant analysis	Commonly used for resequencing and variant detection relative to a reference.	Can support discovery of sequence differences, particularly when reference divergence is high.
Typical use	Model organisms, resequencing, population studies, targeted variant analysis.	Novel organisms, genome projects, highly divergent genomes, genome reconstruction.

PRACTICAL DECISION

Start with the biological question

01 Do you have a high-quality reference genome?

Select an answer to see a general workflow consideration.

05 / LIMITATIONS

Reference Bias: What Can Be Missed?

Reference-based analysis is powerful, but the choice of reference can influence which sequences are successfully mapped, detected, and interpreted.

THE CORE CONCEPT

A reference genome is not always a complete representation of the sample.

In a reference-based workflow, sequencing reads are aligned against a predefined sequence. Reads that differ substantially from the reference may align less efficiently, align incorrectly, or remain unmapped.

As a result, biological variation that is absent or highly divergent from the selected reference can be more difficult to identify.

MAPPING SCENARIO

Where can information be lost?

CLICK THE STEPS

SAMPLE

The sample contains the biological sequence information generated by the sequencing experiment, including variants and sequences that may differ from the selected reference.

POTENTIAL CONSEQUENCES

What types of variation can become difficult to detect?

Highly divergent regions

Reads with substantial sequence divergence from the reference may have reduced or ambiguous alignment.

Insertions

Sequence present in the sample but absent from the reference can be difficult to represent through a conventional reference coordinate system.

Deletions and structural variation

Larger genomic differences can complicate alignment and downstream interpretation.

Novel sequence

Sequence that is not represented in the reference may have no direct reference position to which reads can map.

IMPORTANT DISTINCTION

Reference bias does not mean that reference-based analysis is unreliable.

Reference-based workflows remain highly useful when the reference is appropriate for the organism and analytical question. The limitation is that the reference itself can influence what is represented naturally and what requires additional approaches to detect.

NEXT CONCEPT

What if one reference is not enough?

Modern genomics increasingly uses population-scale representations such as pangenomes to capture genomic diversity that cannot be adequately represented by a single reference sequence.

Explore pangenomes

06 / BEYOND ONE REFERENCE

Beyond a Single Reference: Understanding Pangenomes

A single reference genome provides a useful coordinate system, but it cannot fully represent the genomic diversity of a population or species. Pangenome approaches address this limitation by representing sequence from multiple genomes.

THE SHIFT

01 SINGLE REFERENCE

One reference sequence is used as the primary representation of the genome.

02 PANGENOME

Multiple genomic sequences are represented together to capture shared and variable sequence.

WHAT IS A PANGENOME?

A representation of genomic diversity rather than a single genomic sequence.

A pangenome combines genomic information from multiple individuals or genomes. It can represent regions that are conserved across genomes as well as sequence differences, alternative paths, insertions, deletions, and other forms of variation.

In computational genomics, this information can be represented using structures such as pangenome graphs, which allow alternative genomic paths to be represented within one framework.

CONCEPTUAL MODEL

From a linear reference to a genomic graph

SELECT A MODEL

LINEAR REFERENCE

A B C D E

A linear reference follows one primary sequence path. Alternative sequence may be represented as differences relative to that path.

PANGENOME GRAPH

A graph can represent alternative sequence paths rather than forcing all genomic variation onto one linear sequence.

WHY IT MATTERS

What can a pangenome add to sequencing analysis?

More genomic diversity

Multiple genomes can provide a broader representation of sequence diversity than a single reference.

Alternative sequences

Sequences that differ from the primary reference can be represented directly within the pangenomic framework.

Structural variation

Pangenomic resources can improve representation and analysis of complex structural variation.

Reduced reference bias

A broader reference representation can reduce the limitations associated with relying on a single genomic sequence.

STUDY

EVIDENCE FROM THE HUMAN PANGENOME

A draft human pangenome reference

In a 2023 Nature study, the Human Pangenome Reference Consortium presented a draft pangenome based on 47 phased, diploid assemblies from genetically diverse individuals.

Compared with workflows based on GRCh38, the authors reported improved small-variant discovery accuracy and increased detection of structural variants when analysing short-read data with the draft pangenome.

47 phased diploid assemblies

119M+ additional polymorphic base pairs

104% increase in structural variants detected per haplotype in the reported comparison

A B C D E

FEATURED SCIENTIFIC ARTICLE

A draft human pangenome reference

Liao, W.-W. et al. Nature 617, 312–324 (2023).

The study presents an initial human pangenome resource designed to better capture genomic diversity and overcome limitations of a single reference genome.

Read the original article

Choosing the right strategy for your project

The next step is to connect these concepts to practical project design: reference availability, sequencing type, biological objectives, and computational resources.

Continue

07 / PROJECT DESIGN

Choosing the Right Strategy for Your Project

There is no universally optimal sequencing analysis strategy. The choice depends on the biological question, the availability and quality of a reference, the type of genomic variation being investigated, the sequencing technology, and the computational resources available.

THE CENTRAL QUESTION

What are you trying to learn from the sequencing data?

A reference-based workflow can be highly efficient when a suitable reference genome exists and the main objective is to identify variants relative to that reference. De novo analysis becomes more attractive when a suitable reference is unavailable, incomplete, or when discovering previously unrepresented sequence is itself an objective.

INTERACTIVE FRAMEWORK

Find a suitable starting strategy

STEP 1 OF 4

REFERENCE AVAILABILITY

Is a suitable reference genome available?

Consider whether a reference exists for the organism or population and whether its completeness and quality are appropriate for your project.

BIOLOGICAL OBJECTIVE

What is the primary objective of your analysis?

✓

SUGGESTED STARTING POINT

Reference-based analysis

A reference-based workflow may be appropriate when a suitable reference is available and the primary goal is to identify sequence variation relative to it.

Alignment Variant calling Reference genome

PRACTICAL COMPARISON

Which approach fits your project?

Project characteristic	Reference-based	De novo	Pangenome
High-quality reference available	Strong fit	Possible	Depends on objective
Reference unavailable	Limited	Strong fit	Depends on available genomes
Variant discovery	Strong fit	Possible	Strong for diverse populations
Novel sequence discovery	Limited	Strong fit	Strong
Population diversity	Can be reference-biased	Resource-intensive	Strong fit
Simple downstream workflow	Usually easier	More complex	More computationally demanding

FIVE FACTORS TO CONSIDER

The method should follow the biological question

Reference quality

A fragmented, incomplete, or highly divergent reference can influence mapping and downstream interpretation.

Type of variation

SNPs and small indels may be handled differently from large insertions, deletions, repeats and structural variants.

Sequencing technology

Read length, accuracy, coverage and sequencing platform influence which computational approaches are practical.

Computational resources

Assembly and graph-based analyses can require different levels of memory, storage and computational capacity.

Biological interpretation

The final strategy should support the biological interpretation rather than being selected only for computational convenience.

Sequencing technology changes the analysis

Read length, sequencing accuracy and coverage can strongly influence mapping, assembly and variant detection.

Continue

08 / SEQUENCING TECHNOLOGY

Sequencing Technology Shapes the Analysis

The characteristics of sequencing reads directly influence how genomic data can be aligned, assembled and interpreted. Read length, sequencing accuracy, coverage and platform characteristics should therefore be considered before selecting an analytical strategy.

CORE CONCEPT

The same biological question can require different computational strategies.

Short-read and long-read sequencing technologies provide different types of information. Short reads generally provide high per-base accuracy and can be highly effective for reference-based variant analysis, while longer reads can span repetitive regions and larger genomic events, making them particularly useful for genome assembly and structural variant discovery.

INTERACTIVE COMPARISON

Explore the main sequencing approaches

SHORT READS

SHORT-READ SEQUENCING

High accuracy and efficient reference-based analysis

Short reads can provide highly accurate sequence information and are widely used for applications such as small variant detection, transcript quantification and other reference-oriented analyses.

READ LENGTH Short

PER-BASE ACCURACY High

REFERENCE MAPPING Strong

REPEAT RESOLUTION Limited

LONG-READ SEQUENCING

Longer molecules can resolve complex genomic regions

Long reads can span repetitive or structurally complex regions that are difficult to resolve with shorter fragments. This makes them particularly useful for de novo assembly, structural variant detection and genome characterization.

READ LENGTH Long

PER-BASE ACCURACY Technology-dependent

REFERENCE MAPPING Strong

REPEAT RESOLUTION Improved

HYBRID APPROACHES

Combining complementary sequencing characteristics

Hybrid strategies can combine complementary information from different sequencing technologies. Depending on the project, high-accuracy short reads may complement the long-range information provided by long reads.

READ INFORMATION Complementary

ACCURACY Combined

ASSEMBLY Potentially improved

COMPLEXITY Higher

IMPACT ON ANALYSIS

How read characteristics affect downstream analysis

Mapping

Read length and sequence divergence influence how confidently reads can be aligned to a reference.

Assembly

Longer reads can provide additional connectivity across repetitive and structurally complex regions.

Variant detection

The sequencing platform and read properties influence which classes of variants can be detected reliably.

Coverage

Coverage depth and distribution affect confidence in both variant detection and genome reconstruction.

STRATEGY CONNECTION

Technology and analytical strategy are closely connected

REFERENCE-BASED →

Often efficient when a suitable reference exists and the main objective is comparison against that reference.

Mapping Variant calling Quantification

DE NOVO →

Particularly relevant when reconstructing sequence or discovering genomic content that is not adequately represented by an existing reference.

Assembly Novel sequence Genome reconstruction

From reads to a genome: understanding de novo assembly

The next step is to examine how sequencing reads are transformed into contiguous genomic sequences without relying on a predefined reference.

Continue

09 / DE NOVO ASSEMBLY

From Sequencing Reads to a Reconstructed Genome

De novo assembly attempts to reconstruct genomic sequences directly from sequencing reads without requiring a predefined reference genome. The process relies on relationships between reads to reconstruct longer genomic sequences.

THE ASSEMBLY PROBLEM

Sequencing produces fragments — assembly reconstructs their larger structure.

A sequencing experiment does not normally produce an intact genome sequence. Instead, it generates many reads representing fragments of the original molecules. De novo assembly uses computational methods to infer how these fragments connect and reconstruct longer sequences.

INTERACTIVE PIPELINE

How de novo assembly progresses

STEP 01 / 05

RAW SEQUENCING DATA

Sequencing reads

The process begins with sequencing reads generated from the biological sample. These reads are individual sequence fragments rather than a continuous genome.

ATGCTAGCTAGC GCTAGCTTACGA TTACGATCGGTA CGGTACTGACCT GACCTAGCTTAA

READ RELATIONSHIPS

Finding relationships between reads

Assembly algorithms identify relationships between reads using sequence information. Depending on the assembly strategy, these relationships may be represented using overlap structures or graph-based representations.

ATGCTAGCTAGC

GCTAGCTTACGA

TTACGATCGGTA

CONTIG GENERATION

Building contiguous sequences

Related reads can be assembled into longer contiguous sequences known as contigs. A contig represents a reconstructed stretch of sequence supported by the available reads.

CONTIG

LONGER STRUCTURES

From contigs to scaffolds

When additional information supports the ordering and orientation of contigs, they can be organized into larger structures known as scaffolds. Gaps may remain between individual contigs.

CONTIG A CONTIG B CONTIG C

QUALITY ASSESSMENT

Evaluating the assembly

Assembly should not end when contigs or scaffolds are generated. Quality assessment is required to examine continuity, completeness, accuracy and potential assembly errors.

CONTINUITY How fragmented is the assembly?

COMPLETENESS How much expected sequence is represented?

ACCURACY Are sequence errors present?

ASSEMBLY ALGORITHMS

Different data structures support different assembly strategies

Assembly approaches differ in how they represent the relationships between sequencing reads. Two classical concepts are overlap-based assembly and de Bruijn graph approaches.

01 Overlap-based

Reads are compared to identify sequence overlaps and establish connections between them. This approach is closely associated with overlap-layout-consensus concepts.

READS → OVERLAPS → LAYOUT → CONSENSUS

02 de Bruijn graph

Reads are decomposed into shorter sequence elements called k-mers, which are represented as nodes or edges in a graph depending on the formulation.

READS → K-MERS → GRAPH → ASSEMBLY

KEY CHALLENGES

Why genome assembly is not simply “joining the reads”

Repetitive regions

Repeated sequences can create ambiguous connections between reads and complicate reconstruction.

Sequencing errors

Errors in reads can create false graph connections or incorrect sequence reconstruction.

Heterozygosity

Genetic differences between homologous chromosomes can complicate assembly and produce alternative paths.

Genome complexity

Large genomes and complex genomic structures can substantially increase computational requirements.

Reference bias and what a genome assembly can reveal

Once a reference-based and de novo perspective are understood, the next question is how the choice of reference can influence variant discovery and biological interpretation.

Continue

10 / FINAL COMPARISON

Choosing the Right Strategy for Your Sequencing Data

Reference-based and de novo analysis are not competing methods in every situation. The appropriate strategy depends on the biological question, the availability and quality of a reference genome, sequencing technology, genome complexity and the type of information that needs to be recovered.

THE CENTRAL QUESTION

Do you want to compare your reads to something known, or reconstruct what is present?

This distinction provides a useful starting point when designing a sequencing analysis workflow. In practice, however, many projects benefit from combining multiple analytical approaches.

AT-A-GLANCE COMPARISON

Reference-Based vs De Novo

Criterion	Reference-Based	De Novo
Reference genome	Required	Not required
Main principle	Align reads to an existing reference	Reconstruct sequence from the reads
Typical strength	Efficient comparison and known-region analysis	Discovery of novel or poorly represented sequence
Small variant analysis	Commonly used	Possible, but generally requires additional assembly-aware approaches
Novel sequence discovery	Limited by the reference	Major strength
Computational complexity	Often lower	Often higher
Reference bias	Potential concern	Reduced dependence on a predefined reference
Best suited for	Well-characterized genomes and comparative analyses	Novel genomes, divergent samples and genome reconstruction

PRACTICAL DECISION GUIDE

Which approach should you consider?

Choose reference-based analysis when…

A high-quality reference genome is available.
The objective is to detect differences relative to a known sequence.
Efficient mapping and downstream quantification are priorities.
The biological question focuses on known genomic regions.

Consider de novo analysis when…

No suitable reference genome exists.
The sample may contain novel sequence.
The available reference is highly divergent or incomplete.
Genome reconstruction itself is an objective.

REFERENCE BIAS

A reference is powerful — but it also defines what can be observed.

Reference-based workflows provide an efficient framework for interpreting sequencing reads, but the choice and quality of the reference can influence mapping and downstream interpretation. Highly divergent or absent genomic regions may be difficult to represent adequately when analysis depends entirely on a predefined reference.

De novo approaches provide an alternative perspective by reconstructing sequence directly from the sequencing data. They can therefore complement reference-based analyses when the goal includes discovering sequence that may not be adequately represented by the reference.

BEYOND THE BINARY CHOICE

In many projects, the strongest workflow is not “reference OR de novo”.

Reference-based and de novo approaches can provide complementary information. A project may use assembly to characterize genomic structure and mapping to support precise comparison, annotation or variant analysis.

01 Sequencing Generate reads

→

02 De novo Explore sequence

→

03 Reference Compare & interpret

→

04 Validation Assess findings

KEY TAKEAWAYS

Five principles to remember

A reference-based workflow is generally efficient when a suitable reference exists.

De novo assembly can reveal sequence that is missing or poorly represented in a reference.

Sequencing technology strongly influences which analytical strategies are practical.

Assembly quality must be evaluated rather than judged only by continuity.

The biological question should guide the computational workflow — not the other way around.

SCIENTIFIC REFERENCES

Selected scientific literature

The following publications provide methodological and conceptual foundations for genome assembly, sequencing analysis and reference-based interpretation.

GENOME ASSEMBLY

Velvet: Algorithms for de novo short read assembly

Zerbino & Birney (2008). Genome Research.

Read article ↗

DE BRUIJN GRAPH

ABySS: A parallel assembler for short read sequence data

Simpson et al. (2009). Genome Research.

Read article ↗

ASSEMBLY

Genome assembly in the age of next-generation sequencing

Miller et al. (2010). Briefings in Bioinformatics.

Read article ↗

LONG READS

Canu: scalable and accurate long-read assembly via adaptive k-mer weighting and repeat separation

Koren et al. (2017). Genome Research.

Read article ↗

REFERENCE MAPPING

The Sequence Alignment/Map format and SAMtools

Li et al. (2009). Bioinformatics.

Read article ↗

GENOME ANALYSIS

QUAST: quality assessment tool for genome assemblies

Gurevich et al. (2013). Bioinformatics.

Read article ↗

ARTICLE COMPLETE

Reference-based or de novo?

The answer begins with the biological question, the available reference and the information contained in the sequencing data.

Review de novo assembly ↑